Divide and concer: Data structuring in applicative multiprocessing systems

The conc representation of lists in implementations of applicative languages is introduced. This representation is based on an operator conc which supplements the usual cons, the representation itself being able to coexist with the conventional nested-pair representation. Advantages of the conc representation are presented which relate to multiprocessing and conventional systems. Qualitative comparison with other compact list representations is made. The conc representation exploits demand-driven evaluation and provides further justification for the “tagged data” approach.